***Data preparation***
replace gender =. if gender == 99
label define elite 0 "public" 1 "elite"
label values elite elite
***Figure 1***
logistic strike i.elite i.gender i.party [pw=weight] if scenario == 1
margins i.elite
marginsplot, recast(bar) scheme(s2mono) title("") ytitle("Pr. of selecting nuclear strike") xtitle("") yscale(range(0(0.1)1)) yticks(0(0.1)1) yla(0(0.1)1) name(scat8)
logistic strike i.elite i.gender i.party [pw=weight] if scenario == 2
margins i.elite
marginsplot, recast(bar) scheme(s2mono) title("") ytitle("Pr. of selecting chemical strike") xtitle("") yscale(range(0(0.1)1)) yticks(0(0.1)1) yla(0(0.1)1) name(scat9)
graph combine scat8 scat9, col(2)
***Appendix 4 (regression table)***
logistic strike i.elite i.gender i.party [pw=weight]
estimates store M1
logistic strike i.elite i.gender i.party [pw=weight] if scenario == 1
estimates store M2
logistic strike i.elite i.gender i.party [pw=weight] if scenario == 2
estimates store M3
logistic strike i.scenario i.gender i.party [pw=weight] if elite == 1
estimates store M4
esttab M1 M2 M3 M4 using models.rtf, noeqlines eqlabels(none) eform nogaps se pr2 varlabels(1.elite "Elite" 2.gender "Woman" 2.party "Labour" 3.party "Other party" _cons "Constant") drop(0.elite 1.gender 1.party) mtitle("Dependent variable") title(Logistic regression results) nonumbers mlabels("Model 1" "Model 2" "Model 3" "Model 4")
***Appendix 6 (additional analysis)***
logistic strike i.scenario i.gender i.party [pw=weight] if elite == 1
***Appendix 7 (additional analysis)***
logistic strike i.elite i.gender i.party [pw=weight] 
margins i.elite
marginsplot, recast(bar) scheme(s2mono) title("") ytitle("Pr. of selecting unconventional strike") xtitle("") yscale(range(0(0.1)1)) yticks(0(0.1)1) yla(0(0.1)1) 